首页 | 本学科首页   官方微博 | 高级检索  
     检索      

改进的 Hopfield 网络及汉字识别
引用本文:陈锋.改进的 Hopfield 网络及汉字识别[J].太原科技大学学报,1997(4).
作者姓名:陈锋
作者单位:安徽省教育学院,北京市迪科电子公司
摘    要:针对Hopfield网络的局限性,本文提出了对Hopfield模型及联想记忆算法的改进:(1)改变其权对称性(2)引入反传学习算法(3)为避免在学习过程中出现振荡,引入基本反传值,并作了理论上的分析,运用改进的Hopfield网络进行相近的汉字识别,取得了较好效果。

关 键 词:Hopfield网络  吸引子  联想记忆  基本反传值,模式识别

An improved Hopfield Netwotk and its application to Chinese Character Recognition
Chen Feng.An improved Hopfield Netwotk and its application to Chinese Character Recognition[J].Journal of Taiyuan University of Science and Technology,1997(4).
Authors:Chen Feng
Abstract:A method is pressnted to improve Hopfield Network in this paper.(1)Making weigths of IINN unsymmetrical (2)Back Propagation Learning Algorithm is introduced (3)To avoid oscillating while learning,Basic Backprogation Value is also proposed. The algorthm is also analysed in the paper.When used to recognize Charactcr,the algorithm proves to be quite satisfactory.
Keywords:hopfield network  attractor  pattern recognition  basic backprogation value  assocative memory  
本文献已被 CNKI 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号